def parse(self, response: scrapy.http.response.Response):
base_css = "[data-test=qsp-financial] tbody "
datetimes = response.css(
base_css +
'tr:first-child td:not(:first-child) span::text').extract()
labels = response.css(
base_css +
'tr:not(:first-child) td:first-child:not([colspan]) span::text'
).extract()
values = response.css(
base_css +
'tr:not(:first-child) td:not(:first-child) ::text').extract()
datetimes = list(map(lambda x: x.replace('/', '-'), datetimes))
symbol = response.request.url.split('=')[1]
target_file = os.path.join(YahooFinanceSpider.target_dir,
symbol + '.csv')
current_label_index = -1
current_datetime_index = -1
datetimes_len = len(datetimes)
df = pd.DataFrame(index=labels, columns=datetimes)
pd.options.mode.chained_assignment = None
for i in range(0, len(values)):
current_datetime_index += 1
if i % datetimes_len == 0:
current_label_index += 1
current_datetime_index = 0
val = str(values[i]).replace('-', '')
val = str(val).replace(',', '')
if str(val) != '':
val = int(float(val) *
1000) #TODO check if all numbers are in thousands
df.loc[labels[current_label_index]][
datetimes[current_datetime_index]] = val
mode = 'w'
header = True
if os.path.isfile(target_file):
mode = 'a'
header = False
if df.shape[0] != 0 and df.shape[1] != 0:
with open(target_file, mode) as f:
df.to_csv(f, header=header)
def parse(self, response: scrapy.http.response.Response):
print(response.url)
print('gathering links', response.url)
self.company_links[response.url] = response.css(
'.qTableFull tr td:first-child a::attr(href)').extract()
# Continue only when all company_links are gathered
can_continue = True
for start_url in self.start_urls:
if start_url not in self.company_links:
print('Not all company links yet gathered', response.url)
can_continue = False
break
if can_continue:
print('All links gathered. Proceeding.')
company_links = []
# Organize links in correct order (same as start_urls)
for start_url in self.start_urls:
company_links += self.company_links[start_url]
links_len = len(company_links)
for i, link in enumerate(company_links):
# print(self.url_base + link + self.suffix)
yield scrapy.Request(self.url_base + link + self.suffix,
self.parse_company_page,
priority=links_len - i)
print('Scheduled all requests. Total', links_len)
def parse(self, response: scrapy.http.response.Response):
key_words = ['望京', '望馨花园', '望馨园', '东湖渠']
send = SendEmail()
history = []
with open('history.txt') as f:
tmp = f.readlines()
if len(tmp):
history.extend(tmp)
else:
self.log('历史记录是空', level=logging.WARNING)
page = response.css('td.title')
for i in page:
title = i.css('a::text').extract_first().strip()
link = i.css('a::attr(href)').extract_first()
self.log('租房标题：{0}'.format(title), level=logging.WARNING)
self.log('租房链接：{0}'.format(link), level=logging.WARNING)
email_message = '租房标题：{0}\n租房链接：{1}'.format(title, link)
for j in key_words:
if j in title and link not in history:
# QQ邮箱对发信频率有限制，所以没有找到好的方法之前，无脑 sleep
time.sleep(10)
send.send_email('', email_message)
history.append(link + '\n')
with open('history.txt', 'w') as f:
f.writelines(history)
def parse_list(self, response:scrapy.http.response.Response):
art_links_selector = response.xpath('//*[@id="J_main-container"]//h2[@class="post-title"]/a')
for art_link_selector in art_links_selector:
link = art_link_selector.xpath('@href')
title = art_link_selector.xpath('text()')
# 首页的第二页按钮
second_page = response.xpath('//*[@id="J_main-container"]'
'//a[contains(@class, "home-browser-more-btn")]/@href').get()
if second_page:
yield response.follow(second_page, callback=self.parse_list)
#
next_page = response.xpath('//*[@id="J_main-container"]//ul[@class="pagination"]'
'/li[not(contains(@class, "disabled"))]/a[@aria-label="Next"]/@href').get()
if next_page:
yield response.follow(next_page, callback=self.parse_list)
def parse_ticker_page(self, response: scrapy.http.response.Response):
self.parse_price_page(response)
next_page_href = response.css('.pages_right::attr(href)').extract()
if len(next_page_href) > 0:
time.sleep(1)
return scrapy.Request(self.url_base + next_page_href[0],
self.parse_ticker_page)
def GetMovieCountry(response: scrapy.http.response.Response):
"""
解析制片国
:param response: scrapy返回的response
:return: 制片国家
"""
return response.xpath(
"//div[@id='info']/span[text()='制片国家/地区:'][1]/following-sibling::text()[1]"
).extract_first(default="").strip()
def GetRateInfo(response: scrapy.http.response.Response):
"""
解析评分信息
:param response: scrapy返回的response
:return: 评分人数,评分信息
"""
rateNumber = int(
response.xpath("//a[@class='rating_people']/span/text()").
extract_first(default=0))
rateDetails_dict = dict()
for start_num in range(1, 6):
rate = response.xpath(
"//span[@class='stars{} starstop']/../span[@class='rating_per']/text()"
.format(start_num)).extract_first(default="0")
rateDetails_dict.update(
{start_num: float(rate.strip('%')) / 100.0})
return rateNumber, rateDetails_dict
def parse_subpage(self, response: scrapy.http.response.Response):
links = response.css("#main-content a::attr(href)").extract()
for link in links:
filename = link.rsplit('/', 1)[-1]
pattern = re.compile("^R\d+\.htm$")
if pattern.match(filename):
yield scrapy.Request(self.url_base + link, self.get_data)
def parse(self, response: scrapy.http.response.Response):
next_page = response.xpath(
'//div[@class="navigation-wrapper"]/div/a[@class="next"]/@href'
).get()
if next_page:
print(next_page)
self.count += 1
if self.count < 20:
yield response.follow(next_page, callback=self.parse)
desc = response.xpath('//meta[@name="description"]/@content').get()
tags = response.xpath('//span[@class="tag-links"]/a/text()').getall()
res = self.extractor.extract(response.text)
yield MeituanArticleSpiderItem(url=response.url,
title=res['title'],
content=res['content'],
tags=tags,
author=res['author'],
publish_time=res['publish_time'])
def _parse_role(response: scrapy.http.response.Response) -> Dict[str, str]:
"""
Extract Sphinx role from a crawled page.
Valid roles:
- function
- class
- module
Args:
response: PLACEHOLDER.
Returns:
String containing the role.
"""
url = response.url
name_query = "//h1/text()"
name = response.xpath(name_query).get()
if url in (
"https://www.tensorflow.org/api_docs/python/tf",
"https://www.tensorflow.org/probability/api_docs/python/tfp",
):
return {"name": name, "url": url, "role": "package"}
section_query = "//h2/text()"
sections = response.xpath(section_query).getall()
if "Module" in name.split(": "):
role = "module"
name = name.split(": ")[-1]
elif "Attributes" in sections or "Methods" in sections:
role = "class"
else:
# If the object is not a Module or a Class then it is a function.
role = "function"
return {"name": name, "url": url, "role": role}
def parse(self, response: scrapy.http.response.Response) -> scrapy.Request:
"""
Main scrapy parser
:param response: scrapy response object
:return: new scrapy request
"""
for url in response.xpath(
'//ul[@class="fl_titlelist"]/li/div[@class="fl_name"]/a/@href'
):
url_val = url.extract()
if url_val and url_val.strip('/') in self.already_harvested:
continue
else:
yield scrapy.Request(
url=url_val,
callback=self.parse_item,
cb_kwargs={'on_netflix': '/netflix/' in response.url})
next_url = response.xpath(
'//li[@class="page-item"]/a[text() = "Next"]/@href').extract_first(
)
if next_url:
yield scrapy.Request(url=next_url, callback=self.parse)
def GetActorsInfo(response: scrapy.http.response.Response):
"""
解析演员信息
:param response: scrapy返回的response
:return: 演员信息字典
"""
try:
actor_info_list = response.xpath("//span[@class='actor']//a")
return {
actor_info.xpath("text()").extract_first():
actor_info.xpath("@href").extract_first()
for actor_info in actor_info_list
}
except:
return dict()
def _parse_role(response: scrapy.http.response.Response) -> Dict[str, str]:
"""
Extract Sphinx role from a crawled page.
Valid roles:
- function
- class
- module
Args:
response: PLACEHOLDER.
Returns:
String containing the role.
"""
url = response.url
if response.url == "https://www.tensorflow.org/api_docs/python/tf":
return "package"
name_query = "//h1/text()"
name = response.xpath(name_query).get()
class_selector = response.xpath("//h2/text()").get()
if "Module" in name.split(": "):
role = "module"
name = name.split(": ")[-1]
elif class_selector == "Class ":
role = "class"
else:
# If the object is not a Module or a Class then it is a function.
role = "function"
return {"name": name, "url": url, "role": role}
def GetDirectorOrAuthorInfo(key, response: scrapy.http.response.Response):
"""
解析导演或编剧信息
:param key: '导演'或'编剧'
:param response: scrapy返回的response
:return: 导演或编剧信息字典
"""
try:
info_list = response.xpath(
"//div[@id='info']//span[text()='{key}']/following-sibling::span[1]/a"
.format(key=key))
return {
info.xpath("text()").extract_first():
info.xpath("@href").extract_first()
for info in info_list
}
except:
return dict()
def parse(self, response: scrapy.http.response.Response, **kwargs):
data_list = response.xpath('//*[@id="main-container"]/div[2]/ol/li')
for data in data_list:
item = CcspiderItem()
item['title'] = data.xpath('.//p[1]/text()')[2].get().strip()
item['authors'] = data.xpath('.//p[2]/a/text()').extract()
date = utils.merge_text(data.xpath('.//p[4]/text()[last()]').get())
date = date.split(' ')
item['month'] = utils.month_to_int(date[0])
item['year'] = int(date[1][:4])
item['subjects'] = utils.deduplicate(
data.xpath('.//div/div/span/@data-tooltip').extract())
item['abstract'] = utils.merge_text(
data.xpath('.//p[3]/span[3]/text()').get())
item['citation'] = 0
yield item
print('已爬完{}页（共{}条）'.format(self.page, self.page * self.size))
self.page += 1
def get_data(self, response: scrapy.http.response.Response):
document_type = response.css('th.tl strong').extract_first()
period_label = response.css('th.th::text').extract_first()
dt = response.css('th.th div::text').extract_first()
if period_label is None or document_type is None or dt is None:
# print(period_label)
# print(document_type)
# print(dt)
return
document_type = document_type.lower()
period_label = period_label.lower()
period_labels = ['12 months ended']
document_types = {
'income_statement': 'consolidated statements of income',
'balance_sheet': 'consolidated balance sheets',
'cash_flow': 'consolidated statements of cash flows'
}
is_period_important = False
is_document_important = False
for p_label in period_labels:
if p_label in period_label:
is_period_important = True
break
for slug, d_type in document_types.items():
if d_type in document_type:
is_document_important = True
break
if is_period_important and is_document_important:
if "thousand" in document_type:
multiplier = 1000
elif "million" in document_type:
multiplier = 1000000
elif "billion" in document_type:
multiplier = 1000000000
else:
raise RuntimeError('No multiplier defined in ' + response.url +
'. Document heading: ' + document_type)
year = dt[-4:]
cik = response.url.rsplit('/')[-3]
fin_dict = {'cik': cik}
records = response.css('tr')
for record in records:
record_title = record.css('td.pl a::text').extract_first()
if record_title:
record_title = record_title.replace(',', '')
value = record.css('td.nump::text').extract_first()
# print(record_title, value)
if value:
digit_val = re.findall(r'[\d+,]+', value)[0]
# print('digit_val', digit_val)
if digit_val:
digit_val = float(digit_val.replace(
',', '.')) * multiplier
fin_dict[record_title] = str(digit_val)
file_path = os.path.join(self.output_dir, year + '.csv')
mode = 'w'
if os.path.isfile(file_path):
mode = 'a'
with open(file_path, mode) as f:
print('Saving output to ' + file_path)
#FIXME sort before saving
w = csv.DictWriter(f, fin_dict.keys())
# if mode == 'w':
w.writeheader()
w.writerow(fin_dict)
else:
pass
def parse(self, response: scrapy.http.response.Response):
links = response.css("#main-content a::attr(href)").extract()
for link in links:
yield scrapy.Request(self.url_base + link, self.parse_subpage)
def parse_item(self, response: scrapy.http.response.Response,
on_netflix) -> CritickerMoviesItem:
"""
Extract data from given item url
:param response: scrapy response object
:return: Criticker Movies item object
"""
movie_data = CritickerMoviesItem()
movie_data['on_netflix'] = int(on_netflix)
movie_data['url'] = response.url.strip('/')
movie_data['uid'] = self.extract_uid_from_url(movie_data['url'])
movie_data['type'] = response.xpath(
'//*[@id="fi_info_type"]/text()').extract_first()
movie_data['name'] = response.xpath(
'//h1/span[@itemprop="name"]/text()').extract_first()
movie_data['date_published'] = response.xpath(
'//h1/span[@itemprop="datePublished"]/text()').extract_first()
movie_data['start_date'] = response.xpath(
'//h1/span[@itemprop="startDate"]/text()').extract_first()
movie_data['end_date'] = response.xpath(
'//h1/span[@itemprop="endDate"]/text()').extract_first()
movie_data['poster_url'] = response.xpath(
'//div[@id="poster"]/img/@src').extract_first()
movie_data['description'] = ' '.join([
_.extract().strip()
for _ in response.xpath('//span[@itemprop="description"]//text()')
]).strip()
if not movie_data['description']:
movie_data['description'] = None
more_info_elem = response.xpath('//div[@id="fi_moreinfo"]')
h = more_info_elem.xpath('./p')
for i, hi in enumerate(h):
hi_ = hi.attrib['id']
label = self.extract_label_from_id(hi_)
if 'aka' in label:
movie_data[label] = response.xpath(
'//p[@id="{}"]/text()'.format(hi_)).extract_first()
else:
movie_data[label] = self.extract_more_info(hi)
movie_data['trailer_url'] = response.xpath(
'//div[@id="fi_trailer"]/iframe/@src').extract_first()
if movie_data['trailer_url'] == 'http://www.youtube.com/watch?v=':
movie_data['trailer_url'] = None
movie_data['rss_feed_url'] = response.xpath(
'//*[@id="fi_titlerss"]/a/@href').extract_first()
movie_data['avg_percentile'] = response.xpath(
'//span[@itemprop="ratingValue"]/text()').extract_first()
movie_data['n_ratings'] = response.xpath(
'//span[@itemprop="reviewCount"]/text()').extract_first()
return movie_data
def parse(self, response: scrapy.http.response.Response):
links = response.css(
".list a[href^=\/files\/dera\/data\/financial-statement-data-sets\/]::attr(href)"
).extract()
for link in links:
yield scrapy.Request(self.url_base + link, self.get_data)
